International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248

Volume: 3 Issue: 11 475 – 481

_______________________________________________________________________________________________
Multipath Routing in VANET: Multi-Agent based Approach

Anil D. Devangavi Rajendra Gupta

Information Science and Engg., Basaveshwar Research Cell, AISECT University,
Engineering College, VTU, Madhya Pradesh, India
Bagalkot, India rajendragupta1@yahoo.com
anildevangavi_s@yahoo.co.in

Abstract— In VANET routing of data is a exciting task owing to the high dynamics involved in this network. Delivery of data to the projected
destination turns out to be very puzzling. Single path routing suffers from drawbacks like unreliability and etc. To manage such situation
multipath data delivery is very nominal. In multipath routing more than one path discovered between source and destination node. Data packet
can be sent simultaneously in all paths or data packet can be send by selecting path one after another. It is up to the routing algorithm to select
path thoughtfully to deliver data proficiently. However existing multipath routing protocols even though compute multipath, only one path will
be engaged in actual communication at any given time. Hence this work proposes Multipath Routing in VANET: Multi-agent based Approach
which calculates multiple paths amongst source and destination. Further, all such computed paths will be employed for information
dissemination. NS2 simulation of the proposed approach in realistic mobility models show that it can select more stable link and improve the
network performance.

Keywords- Multipath, Node-disjoint, Congestion.

__________________________________________________*****_________________________________________________
I. INTRODUCTION
Vehicular Ad hoc Network (VANET) is a variety of
network intended for V2V (vehicle-to-vehicle) and V2I
(vehicle to infrastructure) communication It is a core part of
Intelligent Transportation System (ITS) and is being
extensively researched now a days. Unique characteristics of
VANET are no energy resource constraint, high mobility
patterns, relatively high moving speed. During the transit of
information through multi-hop routing, if an intermediate link
was broken, the path needs to be rebuilt or restored. Hence,
efficient routing protocols are essential to function properly.
Multi-path routing protocols aim to compute several paths
during the discovery phase between sender and receiver. Such
network has been extensively used in battle field, disaster
management, under water network etc. Advantages of
multipath routing are 1) Load Balancing 2) Fault Tolerance 3)
Higher Aggregate Bandwidth 4) Lower Delay 5) Reduced
Congestion and improved QoS 6) Higher resiliency 7)
Increased network lifetime [1].
Multipath routing protocols compute node disjoint, link
disjoint or non-disjoint routes in the course of route discovery
Process. Node-disjoint routes have completely disjoint routes
where there are no nodes or links in common. Link-disjoint
routes have no links in common but may have nodes in
common. In an ad hoc network, identification of all node-
disjoint paths between a given pair of nodes is a challenging
task [2].
Rest of the paper is organized as follows. Section I. A
discusses related works. In section I. B methodology is
presented. In section II structure of the proposed work is
discussed. Network environment is mentioned in section II. A.
Section II. B discusses mathematical models. Section II. C
describes the proposed agency. The proposed scheme for the
work is stated in section II. D. Section II. E describes the
algorithm for the proposed work. Simulated model for the
proposed work is discussed in section III. Results and analysis
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
are discussed in section IV. Conclusion is presented in section
V.
A. Related Work
Authors in [3] propose a novel road topology-aware routing
(RTR) protocol for VANETs. To adequately utilize connected
routing paths RTR creates two junction-disjoint paths. RTR
alternately transfers data packets through each established
routing path and dynamically changes the routing path based
on the connectivity of the current path. Authors in [4] propose
Junction-based Multipath Source Routing (JMSR) a geographic
routing protocol which exploits the location of the nodes and
street junctions. Two concurrent paths from the source to the
destination are preserved as a series of junctions the packets
should pass through. Routing information is injected in each
packet. As such every node on the path is aware of the route,
packets to go ahead. Authors in [5] propose distributedly
finding disjoint paths (DFDP) to compute k disjoint paths from
source to destination. Further they analyze the relationship
between disjoint paths and network parameters like robustness,
throughput, and load balancing. Authors in [6] propose link-
disjoint, node-disjoint and zone disjoint multi-path algorithms
aiming on the compromise between lifetime and hop count of
routes for MANET. Authors in [7] evaluate the applicability of
node disjoint multipath in VANET environment. Through
simulation results author„s present that it improves the packet
delivery ratio and reduces the end-to-end delay. Authors in [8]
present a modified form of AOMDV known as Maximally
Spatial Disjoint Multipath routing protocol (MSDM) for
MANETs. The data is communicated employing multiple paths
which are mostly spatially separated and node-disjoint paths.
Results show that spatially node-disjoint routes perform better
than AOMDV which selects only link-disjoint paths.
475
International Journal on Future Revolution in Computer Science & Communication Engineering
Volume: 3 Issue: 11
_______________________________________________________________________________________________
Authors in [9] propose a 2path routing protocol which is a
special case of multipath routing. Two paths are recognized
amongst source and destination. The nodes contributed in
path-1 and path-2 forms a connected graph. Non-interference is
guaranteed as no edge shared between path-1 and path-2.
Authors in [10] present AODV based node disjoint multipath
routing protocol. They focus on residual energy of nodes and
discovery of multiple node disjoint paths with low routing
overhead. Authors in [11] propose a node disjoint multipath
protocol based on AODV with additional features data security
during transmission and link failure. Routing overhead is
substantially reduced because through a single flooding of a
RREQ message route discovery and maintenance process is
completed. Authors in [12] propose a multipath routing
protocol which integrates dynamic clustering and ant colony
optimization. The routing protocol focus on issues related to
energy consumption nodes. The cluster head are elected
considering residual energy. Multiple paths between cluster
head and destination node are computed by ACO. Final
selection of route is based on energy consumption of individual
paths by cluster heads. Network lifetime is thus increased
reducing node energy consumption.
Disadvantages of the existing system:
 Complexity.
 Extra overhead.
 At low bit rate there will be more end-to-end delay.
B) Methodology
In this paper, we propose a new approach for Multipath
Routing in VANET: Multi-agent based Approach. Every
vehicle in the VANET is connected to at least one RSU at any
given time. We use an approach where multiple node disjoint
paths are employed between sender and receiver. The proposed
work thus aims to optimize congestion focusing on multipath
routing. Accordingly information is communicated from sender
vehicle to destination vehicle over multi paths simultaneously.
The proposed scheme works as per the following procedure:
1) Source vehicle identifies destination.
2) Source vehicle triggers mobile agent and creates m clones
of it.
3) The clones (m clones) of the mobile agent thus created
computes neighbouring nodes (next intermediate node) and
migrates to them.
4) The process in step 3 is repeated at each intermediate
vehicle until every clone reaches destination. Multiple paths to
destination are thus calculated.
5) Destination triggers a static agent to compute n node
disjoint paths amongst multiple paths computed in step 4 [15].
6) Destination triggers a static agent to compute the weight
factor of each n path. Weight factor of a path between any two
vehicle nodes is based on the parameters like mobility of
intermediate vehicles in the path, bandwidth utilized, trust
value (behavior) of the intermediate vehicle on the path. It then
prioritizes the n paths based on weight factor with path having
higher weight factor at the top [16].
7) Destination communicates the priority list to the source
vehicle along the best path (path with highest weight factor).
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
ISSN: 2454-4248
475 – 481
8) Source vehicle determines the context of the information to
be sent.
9) If the context is critical, Source communicates the
information on all the paths employing all paths. Else the
information is communicated along the best path to the
destination.
Our contributions are as follows:
 Use of agent technology to compute multipath.
 Use of agents to compute node disjoint paths
 Reduced communication cost owing to the use of
mobile agents
 Various agencies are employed to gather information
of intermediate nodes on the multiple paths and to
prioritize the multiple paths.
The proposed work Multipath Routing in VANET: Multi-
agent based Approach (MRMA) is compared with 2 Path
Routing Protocol (2PR). The advantages of proposed approach
over 2PR are as follows:
 Optimization of congestion.
 All the calculated paths are positively used.
 Higher reliability.
 Competent usage of bandwidth.
II. STRUCTURING THE PROPOSED SSYSTEM
This section presents the network environment, network and
mathematical models, proposed agency, scheme and algorithm.
A. Network Environment
The network structure is depicted in Figure 1.
Figure1. VANET Environment
The above figure presents the network environment in
which we can notice that number of vehicles move with
different speed in different lanes. Some of the assumptions in
this work are: RSU exists at regular intervals and all the RSUs
are networked, each vehicle is connected to at least one RSU at
any given time, all vehicles and RSUs are fitted with GPS. The
location of RSUs is known to one another, Information of all
incoming and outgoing vehicles is updated in the database of
RSU. The database of RSU is updated with information like
current position of vehicle, vehicle‟s mobility and direction.
Each vehicle communicates with other vehicle within its
communication range (VRAN) via V-to-V communication with
road side units via V-to-R communication. The vehicles that lie
within the transmission range of a particular vehicle are the
neighbouring vehicles of that vehicle. Each vehicle‟s status and
information (parameters) is updated by itself and also by
neighbouring vehicles in a local database. The information is
476
International Journal on Future Revolution in Computer Science & Communication Engineering
Volume: 3 Issue: 11
_______________________________________________________________________________________________
related to vehicle ID, position of the vehicle in terms of latitude
and longitude, current speed of the vehicle, direction and etc.
Every vehicle is connected to at least one road side unit (fixed
infrastructure).
B. Network and Mathematical Models
This subsection deals with the network and mathematical
model for the proposed work.
1) Network Model
Link disjoint paths are the path between pair of nodes with no
link in common between the node pair [13]. Node disjoint
paths are path between pair of nodes where no node is
common except source and destination. The node disjoint path
is also link disjoint in nature as no link is common in node
disjoint path between source and destination node. However
link disjointis not always node joint because several links may
share node. In figure 2 three different link disjoint paths shown
between S (source) and D (Destination) node are S-a-b-c-d-D,
S-i-k-e-g-D and S-me-f-D. The last two paths share the node e.
In figure 3 there are two node disjoint paths shown between S
and D node S-a-b-c-d-D and S-e-f-g-h-D. They do not share
any node.
Figure 2.Link Disjoint Path
Figure 3.Node Disjoint Path
In general a path is sequence of vertices: s, v_1.... v_ m, t.
Two paths s, v_1.... v_ m, t and s, u_1.... u_ m, t are called
node-disjoint if v_ i ! = u_ j for any valid I and j.
2) Path weight factor
Path weight factor of each path (Pw) is computed as follows.
Let
MA be the average mobility of the vehicles on the path BAVP be
the bandwidth available on the path and VDEN be the density of
the vehicles on the path Priority of the path (PR) is directly
proportional to bandwidth available on the path (BAVP), the
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
ISSN: 2454-4248
475 – 481
density of the vehicles on the path (VDEN) and inversely
proportional to the average mobility of the vehicles on the path
(MA) (Eq. 1).
Pw BAVP
Pw VDEN
Pw 1 /MA
Pw BAVP * VDEN / MA
Pw = P * BAVP * VDEN / MA (where P is a constant) (1)
C. Proposed Agency
Various agencies are employed to perform communication
among vehicles and road side units viz. Vehicle Agency, Road
Side Unit Agency [14] [15][16] [17].
Vehicle Agency: Vehicle Agency consists of a set of static and
mobile agents like Vehicle Manager Static Agent (VNSA),
Vehicle Manager Knowledge base (VMK) and Path Requester
Agent (PRA).
Vehicle Static Manager Agent: It is a static agent residing
in every vehicle. It is responsible for creating all other static
and mobile agents and knowledge base. It also synchronizes the
agent interactions. It keeps track of parameters like speed,
direction, location, status and unique vehicle ID of the parent
node. It stores the relevant information in VMK. VSMA of
source vehicle triggers mobile agent Multipath Finder Mobile
Agent (MFMA) whenever path is required to destination.
VSMA of the destination triggers static agent (NDPF) to
compute node disjoint paths amongst the multiple paths
computed. VSMA of the destination triggers static agent
(MPE) to assign priorities to each node disjoint path computed.
Figure 4.Vehicle Agency
VSMA of the destination then communicates the prioritized list
of paths to the source along the best path (path with highest
weight factor). It is the responsibility of VSMA to decide on
the no. of paths, based on the context to be used from the
priority list to communicate the information. If the context of
the information is critical then the information is
communicated employing all paths. Else information is
communicated to be source along the best path.
Vehicle Knowledge Base: It is the repository of the vehicle.
All the agents store the relevant information in this repository.
It resembles blackboard and enables agent interactions. It stores
the parameters like speed, direction, location, status and unique
vehicle ID of the parent node and neighbour vehicles.
477
International Journal on Future Revolution in Computer Science & Communication Engineering
Volume: 3 Issue: 11
_______________________________________________________________________________________________
Multipath Finder Mobile Agent (MFMA): It is a mobile agent
triggered by VSMA of source vehicle. Clones of MFMA
computes and migrates to next neighbour vehicles till they
reach destination. Multiple paths are thus created.
Node disjoint Path Finder: It is a static agent triggered by
VSMA of destination vehicle with list of multiple paths as
input parameter. It computes all possible node disjoint paths
amongst the multiple paths.
Multiple Path Evaluator: It is a static agent triggered by
VSMA of the destination. It computes the weight factor of all
the node disjoint paths. Weight factor of a path between any
two vehicle nodes is based on the parameters like mobility of
intermediate vehicles in the path, bandwidth utilized, trust
value (behavior) of the intermediate vehicle on the path. It then
prioritizes the n paths based on weight factor with path having
higher weight factor at the top.
Road Side Unit Agency: It consists of a set of static and
mobile agents viz. Road Side Unit Manager Static Agent
(RMSA), Roadside Unit Knowledge Base RKB and Behavior
Predictor Agent (BPA).
Roadside unit Static Manager Agent: It is a static agent
which resides in every roadside unit. It allocates specific task to
all the created agents and synchronizes the agent interactions. It
monitors the information of incoming vehicles and outgoing
vehicles. The behavior of each outgoing vehicle is
communicated to neighbour RSU.
Figure 5.Road Side Unit Agency
Behavior Predictor Agent (BPA): This is a static agent
residing in every RSU. It keeps track of the behavior (speed,
direction and etc.) of each vehicle in the range of parent RSU.
Using this information it predicts the future behavior of the
vehicle. The relevant information is stored in RKB.
Roadside Unit Knowledge Base (RKB): This is the
repository residing in each RSU. It stores the information of
incoming and outgoing vehicles. It also stores the information
regarding the behavior of each vehicle.
D. Proposed Scheme
 Source identifies destination.
 The VSMA of the source triggers Multipath Finder
Mobile Agent (MPFA) and creates m clones of it.
 The clones (m clones) of MPFA thus created computes
neighbouring nodes (next intermediate node) and
migrates to them [18].
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
ISSN: 2454-4248
475 – 481
 The process in step 3 is repeated at each intermediate
vehicle until every clone reaches destination. Multiple
paths to destination are thus calculated [19] [20].
 The VSMA of destination triggers Node Disjoint Path
Finder (NDPF) to compute n node disjoint paths
amongst multiple paths computed in step 4 [21] [22].
 The VSMA of the destination triggers Multiple Path
Evaluator (MPE). MPE computes the weight factor of
each n path. Weight factor of a path between any two
vehicle nodes is based on the parameters like mobility
of intermediate vehicles in the path, bandwidth utilized,
trust value (behaviour) of the intermediate vehicle on
the path. It then prioritizes the n paths based on weight
factor with path having higher weight factor at the top
[25].
 The VSMA of the destination communicates the
priority list to the source vehicle along the best path
(path with highest weight factor).
 The VSMA of the source vehicle determines the
context of the information to be sent.
 If the context is critical, VSMA of the source
communicates the information on all the paths
employing all paths. Else the information is
communicated along the best path to the destination.
E. Algorithm
Assumptions:
 All the RSUs are networked.
 Each vehicle is connected to atleast one RSU at any
time.
 All vehicles and RSU are fitted with GPS. The location
of RSUs is known to one another.
Algorithm for the communication of information along
multipath to the destination is as below:
Input: A set of vehicles V= {V1…...Vn}
A set of RSUs R= {R1………Rn}
Every vehicle has vehicle agency, RSU has RSU agency.
Let VS: Source Vehicle and VD: Destination Vehicle
DL: Location of destination vehicle
VSMAS: VSMA of source vehicle
VSMAD: VSMA of destination vehicle
Algorithm 1: To Communicate information along
multipath to destination
Begin
VSMAS identifies VD;
VSMAS triggers mobile agent MPMA and creates m clones
of it (DL as input parameter);
Do
At each intermediate vehicle, each clone of MPMA
computes neighbour vehicles and migrates to them;
Until neighbour vehicle = VD
VSMAD triggers NDPF to compute n node disjoint paths;
VSMAD triggers MPE to assign priority (Eq. 1) to each
node disjoint path;
478
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 11 475 – 481
_______________________________________________________________________________________________
VSMAD communicates the priority list to VSMAS along the TABLE I. SIMULATION INPUT PARAMETERS
best path; Simulation parameters Values
VSMAS determine the context of the information; Network simulator ns-2.34
If context = Critical Simulation time Simulation time: 600 seconds
Then Simulation area 5000m X 5000m
Communicate the information to VD employing all paths Number of vehicles 50/100/150
in the list; Communication range 500m
Else Minimum: 20 Km/hr.,
Speed
Communicate the information to VD along the best Maximum: 40 and 60 Km/hr.
path; Data type Constant Bit Rate
End MAC protocol IEEE 802.11e EDCA based DCF
Safety distance between
3mts.
vehicles
III. SIMULATION MODEL Available bandwidth 4000 Mbps
The proposed technique is simulated by taking Bangalore city Road type Free way
map as an example [23]. The simulation is done using NS-2.34
[24] to test the efficacy of approach. We consider “N” number C. Performance Metrics
of vehicles moving in a fixed region of length “A” Km. and
Some of performance metrics evaluated are as follows:
breadth “B” Km. We consider vehicle to move in number of
lanes “L”. Communication coverage area for each vehicle is TABLE II.PERFORMANCE METRICS
considered as “VRAN” meters. At the start of the simulation, Metric Description
vehicles are uniformly distributed in lanes. We assume free Packet Delivery It is the ratio of no. of packets reaching the destination to
flow movement of vehicles and congestions are ignored. Every Ratio (PDR) the no. of packets originating from the source node.
Path Discovery It is the time taken to compute multipath to destination
vehicle is presumed to be equipped with a communication
Time from source.
device and knows start position, start time of vehicle, route that
Transmission It is the time spent by the data packet to reach destination
it selects, and speed at which it travels. Safety distance of “R” Time from the source.
meters between vehicles is assumed. It is the amount of time spent in the dissemination of
Communication
added information other than intended information.
Overhead
Overhead is due to the formation of multipath.

IV. SIMULATION RESULTS AND ANALYSIS

This section presents the results obtained during simulation.
We compare results of proposed work with an existing
multipath routing 2PR. The below mentioned figures are
generated based on the simulation results.

Transmission Time
Figure 7 shows the Transmission Time evaluated for 2PR
and MPMA protocols by increasing the number of paths
between the source and destination in urban scenario. From
Figure 6.Bangalore City Map the figure it is evident that the transmission time decreases
with the increase in the number of paths. Nevertheless the
A. Simulation Procedure
transmission time of MPMA is lower than 2PR routing
Simulation inputs are as follows: A= 5000m, B= 5000m, protocol. This is because as the number of vehicles increases
N= 50, VRAN = 300m, I= 20 Kmph, J= 60 Kmph, L= 2, R = 4 and also due to the multipath between the source and
mts. Simulation procedure for proposed intelligent agent model destination the stability is more. Because of which
is as follows. retransmissions are minimized.
Begin
 Generate VANET in given road length by placing
vehicles uniformly.
 Maintain a data structure at each vehicle to store
information as detailed by scheme.
 Apply mobility to nodes.
 Generate agency.
 Compute performance of system.
End

B. Simulation Inputs
The simulation input parameters are as below: Figure 7.Transmission time V/s No. of vehicles

479
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
International Journal on Future Revolution in Computer Science & Communication Engineering
Volume: 3 Issue: 11
_______________________________________________________________________________________________
Packet Delivery Ratio (PDR)
Figure 8 shows the Packet Delivery ratio (PDR) evaluated
for 2PR and MPMA protocols by increasing the mobility of
the vehicles in the urban scenario. Number of paths is taken as
the third quadrant. From the figure it is evident that PDR
increases with the increase in the number of paths despite
increase in the mobility of vehicles. The PDR of MPMA is
more than 2PR routing protocol. This is because the intelligent
software agents follow the defined trajectory, stability of the
multipath.
Figure 8.Packet Delivery Ratio V/s Mobility
Communication Overhead
Figure 9 shows the communication overhead evaluated for
2PR and MPMA protocols by increasing the mobility in the
urban scenario. Due to the computation of multipath and other
related activities the overhead increases with the increase in
the mobility as illustrated in the figure. The communication
overhead of MPMA is lower than 2PR routing protocol even
at higher mobility. This is because the proposed work
computes multipath employing intelligent agents which
migrate carrying minimum information.
Figure 9.Communication overhead V/s No. of vehicles
V. CONCLUSION
VANETs unique features such as high speed nodes,
continuously changing topology and etc. make them different
from MANETs. With many of the dominant multipath routing
protocols, all the computed paths are not operated. Therefore,
in this paper, we have proposed an efficient multipath routing
protocol called Multipath Routing in VANET: Multi-agent
based Approach. The proposed approach works on all
calculated multipath to transfer the information. The approach
is tested by comparing its performances with 2 Path Routing
(2PR). From the results shown in last section, we can notice
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________
ISSN: 2454-4248
475 – 481
that our approach performs better w.r.t transmission time,
number of multipath computed, communication overhead and
packet delivery ratio. The simulation results show better
performance achieved by MPMA with multipath routing. In
future work, we foresee the process of allocation of granularity
which specifies the smallest unit of information allocated to
each path.
REFERENCES
[1] Jing Wu and Yuhao Wang “Performance Study of Multi-Path in
VANETs and Their Impact on Routing Protocols”, Science
Research on Wireless Engineering and Technology, July 2011,
vol. 2, pp. 125-129, doi:10.4236/wet.2011.23018.
[2] S. S. Dorle, Bhushan Vidhale, and Megha Chakole,
“Evaluation of Multipath, Unipath and Hybrid Routing
Protocols for Vehicular Ad Hoc Networks”, Proc. IEEE
International Conference on Emerging Trends in Engineering &
Technology, Nov. 2011, Mauritius, pp. 311 - 316, doi:
10.1109/ICETET.2011.66.
[3] Hongli Zhang and Qiang Zhang, “A Novel Road Topology-
aware routing in VANETs”, vol. 8710, APWeb 2014: Web
Technologies and Applications, Sept. 2014, pp. 167-176.
[4] Pavlos Sermpezis, Georgios Koltsidas, and Fotini-Niovi
Pavlidou, “Investigating a Junction-Based Multipath Source
Routing Algorithm for VANETs”, IEEE Transl. J. Magn,
March 2013, vol. 17, pp. 600 – 603, doi:
10.1109/LCOMM.2013.011513.122045.
[5] Kejia Zhang, Guisheng Yin, Qilong Han, and Junyu Lin,
“DFDP: A Distributed Algorithm for Finding Disjoint Paths in
Wireless Sensor Networks with Correctness Guarantee”,
International Journal of Distributed Sensor Networks, Jan.
2014, vol. 1, pp. 1-11, doi.org/10.1155/2014/258959.
[6] Natarajan Meghanathan,“ Performance Comparison of Link,
Node and Zone Disjoint Multi-Path Routing Strategies and
Minimum Hop Single Path Routing for Mobile Ad Hoc
Network”, International Journal of Wireless & Mobile
Networks (IJWMN), Nov. 2010, vol. 2, pp. 13-19,
doi:10.5121/ijwmn.2010.2402.
[7] Xiaoxia Huang, and Yuguang Fang, “Performance Study of
Node-Disjoint Multipath Routing in Vehicular Ad Hoc
Networks”, IEEE Transactions on Vehicular Technology, Oct.
2009, vol. 58, pp. 1942-1950, doi:
10.1109/TVT.2008.2008094.
[8] Wesam Almobaideen, Roba Al-Soub, and Azzam Sleit,
“MSDM: Maximally Spatial Disjoint Multipath Routing
Protocol for MANET”, Communications and Network, Nov.
2013, vol. 5, pp. 316-322, doi: 10.4236/cn.2013.54039.
[9] Abhishek Bande and Gaurav Deshmukh, “Node Disjoint
Multipath Routing Approach for Controlling Congestion in
MANET‟s ”, Global Journal of Computer Science and
Technology Network, Web & Security, 2012, vol.12, pp. 39-45.
[10] V.K. Barbudhe et al., “Node-disjoint Multipath Routing
Method Based on AODV Protocol”, International Journal of
Research in Advent Technology (IJRAT), Feb. 2014, vol. 2, pp.
1-6.
[11] Shuchita Upadhayaya and Charu Gandhi, “Node Disjoint
Multipath Routing Considering Link and Node Stability
protocol: A characteristic Evaluation”, International Journal of
Computer Science Issues (IJCSI), Jan. 2010, vol. 7, pp. 18-25.
[12] Jing Yang, Wei Zhao, Mai Xu and Baoguo Xu, “A Multipath
Routing Protocol Based on Clustering and Ant Colony
Optimization for Wireless Sensor Networks”, International
Journal Computer Network and Information Security (IJCNIS),
May 2010, vol. 1, pp. 49-59, doi:10.3390/s100504521.
[13] Indrani Das , D. K. Lobiyal and C. P. Katti “An Analysis of
Link Disjoint and Node Disjoint Multipath Routing for Mobile
Ad Hoc Network”, I. J. Computer Network and Information
480
International Journal on Future Revolution in Computer Science & Communication Engineering ISSN: 2454-4248
Volume: 3 Issue: 11 475 – 481
_______________________________________________________________________________________________
Security (IJCNIS), March 2016, pp. 52-57,
doi:10.5815/ijcnis.2016.03.07.
[14] Samira Harrabi, Walid Chainbi and Khaled Ghedira, “A Multi-
Agent Proactive Routing Protocol for Vehicular Ad-Hoc
Networks”, Proc. IEEE International Symposium on Network,
Computers and Communications, July 2014, Tunisia, pp. 1-6,
doi: 10.1109/SNCC.2014.6866523.
[15] Balamurugan G, S. Punniakodi, Arulalan V. and C.
Jayanavithraa, “Multipath Forwarding in VANET for Message
Dissemination”, International Journal of Advanced Research in
Computer and Communication Engineering, Nov. 2013, vol. 2,
pp. 4188-4191
[16] Dr. A.V. Sutagundar and Dr. S. S. Manvi, “Location aware
event driven multipath routing in Wireless Sensor Networks:
Agent based approach”, Science Direct Egyptian Informatics
Journal, March 2013, vol. 14, pp. 55–65,
doi.org/10.1016/j.eij.2013.01.003.
[17] Hridya C. Premkumar, Vijayashree R. Budyal, Mahabaleshwar
S. Kakkasageri , “Cognitive Agent based Stable Routing
Protocol for Vehicle-to-Vehicle Communication”, Proc. IEEE
Annual India Conf. (INDICON), Feb 2017, pp. 1-5,
doi:10.1109/INDICON.2016.7839102.
[18] Junichi Aoki, Takeshi Hashimoto, T. Ohta and Yoshiaki
Kakuda, “Experimental Evaluation of Location-based Mobile
Agent Migration Mechanism for Information Dissemination
Scheme in VANETs”, Proc. IEEE Symposium on Computing
& Networking (CANDAR), March 2015, pp. 271-274, doi:
10.1109/CANDAR.2015.51.
[19] Dr. A. V. Sutagundar and Dr. S. S. Manvi, “Wheel Based Event
triggered Data Aggregation & Routing in Wireless Sensor
Networks: Agent Based Approach”, Wireless Personal
Communications (Wireless Pers Commun), July 2013, vol. 71,
pp. 491-517, doi: 10.1007/s11277-012-0825-x.
[20] Harish Laxmichand Sharma and Pankaj Agrawal, “Multipath
Reliable Range Node Selection Distance Vector Routing for
VANET: Design Approach”, Proc. IEEE Conf. Electronic
Systems, Signal Processing and Computing Technologies, Feb.
2014, pp. 280-283, doi: 10.1109/ICESC.2014.53.
[21] Xiaoxia Huang, and Yuguang Fang, “Performance Study of
Node-Disjoint Multipath Routing in Vehicular Ad Hoc
Networks”, IEEE Transl. on Vehicular Technology, May 2009,
vol. 58, pp. 1942-1950, doi: 10.1109 /TVT.2008.2008094.
[22] Mahesh K. Marina and Samir R. Das, “On-demand Multipath
Distance Vector Routing in Ad Hoc Networks “, Proc. IEEE
Conf. on Network Protocols, Nov. 2001, pp. 14-23,
doi: 10.1109/ICNP.2001.992756.
[23] Bangalore City Map, Available: www.mapofbangalore.com
[November 2007].
[24] Network simulator: ns-2, Available at:
http://www.isi.edu/nsnam/ns [October 2010].
[25] Mehmet Kurt, Murat Berberler, and Onur Ugurlu, “A New
Algorithm for Finding Vertex-Disjoint Paths”, The
International Arab Journal of Information Technology (IAJIT),
Nov. 2015, vol. 12, pp. 550-555.

481
IJFRCSCE | November 2017, Available @ http://www.ijfrcsce.org
_______________________________________________________________________________________